Lightweight, self-hosted sandbox for AI coding agents that protects against prompt-injected or misbehaving agents: all egress traffic is TLS-inspected and secret-scanned, and credentials are injected at the proxy so the agent never sees them. No third-party platform in the loop, no trust required.
Updated 2026-06-25 22:19:41 -04:00
Updated 2026-06-19 04:20:07 -04:00